Obtaining Thickness Maps of Corneal Layers Using the Optimal Algorithm for Intracorneal Layer Segmentation.

Area of Science:

• Ophthalmology
• Medical Imaging
• Biomedical Engineering

Background:

• Optical Coherence Tomography (OCT) is crucial for diagnosing corneal diseases by providing cross-sectional eye images.
• Manual segmentation of OCT images for corneal thickness mapping is time-consuming and lacks precision.
• Accurate corneal layer segmentation is essential for diagnosis and treatment planning.

Purpose of the Study:

• To develop and evaluate automatic segmentation methods for clinically important corneal layer boundaries on OCT images.
• To create corneal thickness maps using automated segmentation.
• To compare the accuracy of different segmentation methods against manual segmentation by specialists.

Main Methods:

• Applied Gaussian Mixture Model (GMM), Graph Cut, and Level Set methods for automatic segmentation of corneal layers in OCT images.
• Generated three-dimensional corneal data and derived layer thickness maps.
• Calculated mean and standard deviation of corneal layer thickness in various zones for normal subjects.
• Validated segmentation accuracy by comparing automated results with manual segmentation by two corneal specialists.

Main Results:

• The Gaussian Mixture Model (GMM) demonstrated superior accuracy in segmenting corneal layer boundaries compared to Graph Cut and Level Set methods.
• Automated segmentation enabled the creation of detailed corneal thickness maps.
• Quantitative analysis showed GMM as the most accurate method for boundary segmentation.

Conclusions:

• Automatic segmentation of corneal layers using OCT images is feasible and accurate, with GMM showing the best performance.
• This automated approach offers a more efficient and precise alternative to manual segmentation for corneal analysis.
• The developed methods can aid in the diagnosis and treatment of corneal diseases through accurate thickness mapping.